Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > sci.physics > #533131 > unrolled thread

Re: Floating Point - Ping F. Russel

Back to article view | Back to sci.physics

Contents

  Re: Floating Point - Ping F. Russel root <root@localhost.localdomain> - 2015-11-14 21:53 +0000
    Re: Floating Point - Ping F. Russel Fabian Russell <root@localhost.localdomain> - 2015-11-14 22:35 +0000
      Re: Floating Point - Ping F. Russel noTthaTguY <abu.kuanysh05@gmail.com> - 2015-11-14 15:23 -0800
      Re: Floating Point - Ping F. Russel jimp@specsol.spam.sux.com - 2015-11-14 23:29 +0000
        Re: Floating Point - Ping F. Russel Fabian Russell <root@localhost.localdomain> - 2015-11-15 00:00 +0000
          Re: Floating Point - Ping F. Russel jimp@specsol.spam.sux.com - 2015-11-15 01:04 +0000

#533131 — Re: Floating Point - Ping F. Russel

On Sat, 14 Nov 2015 21:24:50 +0000, Lester Jones IV wrote:

> Drop this into your tests.  Everybody is doing it.
> 
> http://lipforge.ens-lyon.fr/www/crlibm/index.html
> 
> C'est la vie.
>

Hey, thanks Podolsky.

Yeah.  Everybody except the wise-asses on this NG.  But they're all just
do-nothing technician types, you know, the glorified "new age" secretary.
Most of them never had a creative thought in their entire lives.

But hey, I found a decent new book: "Handbook of Floating Point Arithmetic"
by Jean-Michel Muller and a slew of others, with most being in France and Europe.
I guess the Amerikans can't tell the trees from the forest, if you know what
I mean.  Anyway it's ISBN 978-0-8176-4704-9 (at the usual place).

They mention the UCBTEST on page 115.  But they don't mention that it
needs patching.  Oh well, if I can figure it out, anyone can.

Just for the heck of it, I'm now coding the Payne/Hanek reduction
in Intel assembly for the 80-bit stack.  I'll let you know that pans out.

Oh yeah, you'd better watch out for these NG freaks.  They're more interested
in headers than physics.  I guess it's the only thing they can really
handle.  LOL!

Ciao

[toc] | [next] | [standalone]

#533143

On Sat, 14 Nov 2015 21:53:57 +0000, root wrote:

> 
> But hey, I found a decent new book: "Handbook of Floating Point Arithmetic"
> by Jean-Michel Muller and a slew of others, with most being in France and Europe.
> I guess the Amerikans can't tell the trees from the forest, if you know what
> I mean.  Anyway it's ISBN 978-0-8176-4704-9 (at the usual place).
> 

Hey, Podolsky.  I barely get past two pages in this book and I'm already
bowled over.

Back in the 1950's the Russkies (Soviets) developed a ternary computer
called Setun.  That's TERNARY and not binary or decimal.

Very strange, one would think.

But it seems that it's a fantastic idea. For one thing, rounding
to nearest is equivalent to truncation!  Floating point efficiency
is also maximized, at a constant MAX_FP, by using a radix of 3.

Also check this out from http://www.computer-museum.ru/english/setun.htm

"Simplicity, economy and elegance of computer architecture are the direct
and practically very important consequence of the ternarity, more exactly
– of representation of data and instructions by symmetrical (balanced) code,
i.e. by code with digits 0, +1, -1. In opposite to binary code there is no
difference between "signed" and "unsigned" number. As a result the amount of
conditional instructions is decrease twice and it is possible to use them more
easily; the arithmetic operations allow free variation of the length of operands
and may be executed with different lengths; the ideal rounding is achieved simply
by truncation, i.e. the truncation coincides with the rounding and there is the
best approximation the rounding number by rounded."

It seems that those Russkies aren't the backward peons the propaganda of
the West makes them out to be.

Shoot!  I should have gone to school in Moscow.

Ciao.

[toc] | [prev] | [next] | [standalone]

#533152

same as truncating -- easy to check;
there are no even numbers per se,
that you can tell bay a final 2 or 0

> But it seems that it's a fantastic idea. For one thing, rounding
> to nearest is equivalent to truncation!  Floating point efficiency
> is also maximized, at a constant MAX_FP, by using a radix of 3.
> 
> Also check this out from http://www.computer-museum.ru/english/setun.htm
> 
> "Simplicity, economy and elegance of computer architecture are the direct
> and practically very important consequence of the ternarity, more exactly
> - of representation of data and instructions by symmetrical (balanced) code,
> i.e. by code with digits 0, +1, -1. In opposite to binary code there is no
> difference between "signed" and "unsigned" number. As a result the amount of
> conditional instructions is decrease twice and it is possible to use them more
> easily; the arithmetic operations allow free variation of the length of operands
> and may be executed with different lengths; the ideal rounding is achieved simply
> by truncation, i.e. the truncation coincides with the rounding and there is the
> best approximation the rounding number by rounded."
> 
> It seems that those Russkies aren't the backward peons the propaganda of
> the West makes them out to be.
> 
> Shoot!  I should have gone to school in Moscow.
> 
> Ciao.

[toc] | [prev] | [next] | [standalone]

#533160

Fabian Russell <root@localhost.localdomain> wrote:
> On Sat, 14 Nov 2015 21:53:57 +0000, root wrote:
> 
>> 
>> But hey, I found a decent new book: "Handbook of Floating Point Arithmetic"
>> by Jean-Michel Muller and a slew of others, with most being in France and Europe.
>> I guess the Amerikans can't tell the trees from the forest, if you know what
>> I mean.  Anyway it's ISBN 978-0-8176-4704-9 (at the usual place).
>> 
> 
> Hey, Podolsky.  I barely get past two pages in this book and I'm already
> bowled over.

Hardly surprising.

> Shoot!  I should have gone to school in Moscow.

You would have been shot.


-- 
Jim Pennino

[toc] | [prev] | [next] | [standalone]

#533171

On Sat, 14 Nov 2015 23:29:07 +0000, jimp wrote:

> 
>> Shoot!  I should have gone to school in Moscow.
> 
> You would have been shot.
>

Ooops!  I forgot.  This guy is a COLD WARRIOR.

Well, the cold war is long over, and so are you.

[toc] | [prev] | [next] | [standalone]

#533183

Fabian Russell <root@localhost.localdomain> wrote:
> On Sat, 14 Nov 2015 23:29:07 +0000, jimp wrote:
> 
>> 
>>> Shoot!  I should have gone to school in Moscow.
>> 
>> You would have been shot.
>>
> 
> Ooops!  I forgot.  This guy is a COLD WARRIOR.
> 
> Well, the cold war is long over, and so are you.
 
Ooops!  I forgot.  This guy is stuck in the 1970's and hasn't learned
a thing since.
 

-- 
Jim Pennino

[toc] | [prev] | [standalone]

Back to top | Article view | sci.physics

csiph-web